Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

I love this interview with Sam Whitmore because it weaves in agentic memory, continual learning, user goal alignment, predictions for what's next... with the playfulness and whimsy of growing a young family while staying on the frontier of tech 🤍 a reminder that Cursor is hiring, & to reach...

ashebytes's profile picture

ashe

10,809 subscribers

13,222 Aufrufe • vor 4 Monaten •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

On agent factories, memory, continual learning & building on the frontier while growing a family In conversation with Sam Whitmore Sam Whitmore, eng at Cursor this is a special one! dive in: 00:25 Sam’s background 01:38 Family building while in frontier tech 11:50 Agent factories at cursor 15:46 Continual learning, agent self-awareness, memory, and user goal alignment 23:51 Building a thoughtful product culture at Cursor 27:00 Thinking about defensibility in 2026 30:01 Selling outside of the tech/X bubble 33:49 Kids & AI use 35:12 On reading fiction 38:32 On pregnancy, hormones, and building 42:58 On agent predictions for the next 6 months 46:35 Hiring at Cursor
47:58

On agent factories, memory, continual learning & building on the frontier while growing a family In conversation with Sam Whitmore Sam Whitmore, eng at Cursor this is a special one! dive in: 00:25 Sam’s background 01:38 Family building while in frontier tech 11:50 Agent factories at cursor 15:46 Continual learning, agent self-awareness, memory, and user goal alignment 23:51 Building a thoughtful product culture at Cursor 27:00 Thinking about defensibility in 2026 30:01 Selling outside of the tech/X bubble 33:49 Kids & AI use 35:12 On reading fiction 38:32 On pregnancy, hormones, and building 42:58 On agent predictions for the next 6 months 46:35 Hiring at Cursor

ashe

59,539 Aufrufe • vor 4 Monaten

challenged myself to make a marketing video for Dot by New Computer in <2 hours. i’ve been a heavy power user since their launch and wanted to constrain myself to storyboard, film and edit in one session. tag and Sam Whitmore so they see this and feel the love !
0:24

challenged myself to make a marketing video for Dot by New Computer in <2 hours. i’ve been a heavy power user since their launch and wanted to constrain myself to storyboard, film and edit in one session. tag and Sam Whitmore so they see this and feel the love !

steve

14,434 Aufrufe • vor 2 Jahren

There is no better place on the internet to learn about Context Engineering than this repo. It's literally a course with a learning path. It gathers the best resources and covers the theory + code for anything related to context, RAG, memory, and agentic systems and more.
0:26

There is no better place on the internet to learn about Context Engineering than this repo. It's literally a course with a learning path. It gathers the best resources and covers the theory + code for anything related to context, RAG, memory, and agentic systems and more.

ℏεsam

48,999 Aufrufe • vor 1 Jahr

Kimmy Granger revealed that she is happy she entered the advlt entertainment industry at a young age because it gave her more time to grow, learn, and gain experience throughout her twenties. She explained that she even wished she had started earlier and encouraged young women entering the industry to approach it as a learning experience and grow with it over time. “I’m really glad that I got in when I was young because I could use all of my twenties growing and learning and expanding in the industry.”
2:12

Sensitive content

Kimmy Granger revealed that she is happy she entered the advlt entertainment industry at a young age because it gave her more time to grow, learn, and gain experience throughout her twenties. She explained that she even wished she had started earlier and encouraged young women entering the industry to approach it as a learning experience and grow with it over time. “I’m really glad that I got in when I was young because I could use all of my twenties growing and learning and expanding in the industry.”

ZAN

231,947 Aufrufe • vor 2 Monaten

【FFXIV】[SAM]vfx MOD：ONIKAGE SHURA Staying true to the essence of the original SAM, we have reimagined it with a wilder, more aggressive presence and overwhelming visual impact. Support us:
4:31

【FFXIV】[SAM]vfx MOD：ONIKAGE SHURA Staying true to the essence of the original SAM, we have reimagined it with a wilder, more aggressive presence and overwhelming visual impact. Support us:

老钱_Laoqian_

67,052 Aufrufe • vor 2 Monaten

New short course: Long-Term Agentic Memory with LangGraph. Learn to build an agent with long-term memory in this course developed in collaboration with taught by its Co-Founder and CEO, Harrison Chase! Personal assistance and productivity tasks have become important use cases for agents. An important feature of an AI assistant, such as a coding or calendar assistant, is its ability to keep improving over time from its experience. Agent memory is the key capability that enables this. To add memory to an agent, you must first figure out what to store and what to retrieve when it is time to use the information. Additionally, you’ll have to decide when to update the stored information. For example, you might update in each iteration loop of the agent or perform updates in the background, with a helper agent. In this course, you will learn a mental framework to build agents with long-term memory. You'll create a useful email assistant that can respond, ignore, and notify using writing, scheduling, and memory-management tools. You’ll develop your agent's memory by adding facts to its memory store, provide examples to learn the user's preferences, and optimize system prompts to evolve instructions based on previous responses. In detail, you’ll: - Learn how the three types of memory--semantic, episodic, and procedural–and the two update mechanisms–via hot path and in the background–apply to your agents. - Build an email agent with writing, scheduling, and availability tools, along with a router that triages incoming email and handles it accordingly by ignoring, responding, or notifying the user. - Add tools to your email agent that allow it to operate on semantic memory by learning facts about the user, storing them in a long-term memory store, and searching over them in future interactions. - Incorporate episodic memory, in the form of few-shot examples, in the triage step of your agents to help them learn and update user preferences. - Add procedural memory as system prompts, optimized with feedback to improve the instructions the agent follows. Learn how to approach memory in agents, and start building agents with long-term memory with LangGraph! Please sign up here:
2:06

New short course: Long-Term Agentic Memory with LangGraph. Learn to build an agent with long-term memory in this course developed in collaboration with taught by its Co-Founder and CEO, Harrison Chase! Personal assistance and productivity tasks have become important use cases for agents. An important feature of an AI assistant, such as a coding or calendar assistant, is its ability to keep improving over time from its experience. Agent memory is the key capability that enables this. To add memory to an agent, you must first figure out what to store and what to retrieve when it is time to use the information. Additionally, you’ll have to decide when to update the stored information. For example, you might update in each iteration loop of the agent or perform updates in the background, with a helper agent. In this course, you will learn a mental framework to build agents with long-term memory. You'll create a useful email assistant that can respond, ignore, and notify using writing, scheduling, and memory-management tools. You’ll develop your agent's memory by adding facts to its memory store, provide examples to learn the user's preferences, and optimize system prompts to evolve instructions based on previous responses. In detail, you’ll: - Learn how the three types of memory--semantic, episodic, and procedural–and the two update mechanisms–via hot path and in the background–apply to your agents. - Build an email agent with writing, scheduling, and availability tools, along with a router that triages incoming email and handles it accordingly by ignoring, responding, or notifying the user. - Add tools to your email agent that allow it to operate on semantic memory by learning facts about the user, storing them in a long-term memory store, and searching over them in future interactions. - Incorporate episodic memory, in the form of few-shot examples, in the triage step of your agents to help them learn and update user preferences. - Add procedural memory as system prompts, optimized with feedback to improve the instructions the agent follows. Learn how to approach memory in agents, and start building agents with long-term memory with LangGraph! Please sign up here:

Andrew Ng

131,850 Aufrufe • vor 1 Jahr

OpenAI's Chief Scientist, Jakub Pachocki, on the continual learning wave: frontier labs are already building this into the core of the technology. The entire premise of scaling was to create systems that learn in context. Jakub says continual learning is not some separate missing piece, but “exactly what we’re working toward.”
0:52

OpenAI's Chief Scientist, Jakub Pachocki, on the continual learning wave: frontier labs are already building this into the core of the technology. The entire premise of scaling was to create systems that learn in context. Jakub says continual learning is not some separate missing piece, but “exactly what we’re working toward.”

Jacob Effron

15,065 Aufrufe • vor 4 Monaten

🔉 Introducing SAM Audio, the first unified model that isolates any sound from complex audio mixtures using text, visual, or span prompts. We’re sharing SAM Audio with the community, along with a perception encoder model, benchmarks and research papers, to empower others to explore new forms of expression and build applications that were previously out of reach. 🔗 Learn more:
1:09

🔉 Introducing SAM Audio, the first unified model that isolates any sound from complex audio mixtures using text, visual, or span prompts. We’re sharing SAM Audio with the community, along with a perception encoder model, benchmarks and research papers, to empower others to explore new forms of expression and build applications that were previously out of reach. 🔗 Learn more:

AI at Meta

1,251,615 Aufrufe • vor 8 Monaten

I'm 25, turning 26 this month. I switched from physics to tech just a few months ago. I started learning programming with no connections in tech - and yes, of course, I often feel lost, alone, and behind. But as ThePrimeagen said, for people like us, there’s more reason and more pressure to learn. I still don’t know if this was the right move, but in the past 7–8 months, I’ve learned a lot - and more importantly, I’ve enjoyed it, something I rarely felt while doing anything else. All of it happened because I stayed consistent, even while doing odd jobs. Now, my next goal is to land a job or a paid internship - just something that can cover my expenses so I can go full-time in tech and keep learning by working on real projects in the industry. So, if you’ve started late or felt lost, watch this... Thank you ThePrimeagen for this❤️
1:37

I'm 25, turning 26 this month. I switched from physics to tech just a few months ago. I started learning programming with no connections in tech - and yes, of course, I often feel lost, alone, and behind. But as ThePrimeagen said, for people like us, there’s more reason and more pressure to learn. I still don’t know if this was the right move, but in the past 7–8 months, I’ve learned a lot - and more importantly, I’ve enjoyed it, something I rarely felt while doing anything else. All of it happened because I stayed consistent, even while doing odd jobs. Now, my next goal is to land a job or a paid internship - just something that can cover my expenses so I can go full-time in tech and keep learning by working on real projects in the industry. So, if you’ve started late or felt lost, watch this... Thank you ThePrimeagen for this❤️

Sid

183,039 Aufrufe • vor 9 Monaten

Text-to-SQL is dead 💀 The next generation of querying is agentic - and it’s already here. This paper ( introduces a type of agentic querying called 𝗙𝘂𝗻𝗰𝘁𝗶𝗼𝗻 𝗖𝗮𝗹𝗹𝗶𝗻𝗴 that uses an LLM to structure queries using predefined function calls in JSON format, with optional arguments for search, filters, aggregation, and grouping. Along with that, it also tests out a bunch of different models with a new dataset, DBGorilla, designed to evaluate agentic querying techniques on real-world use cases. Weaviate AI Database also just released a 𝗤𝘂𝗲𝗿𝘆 𝗔𝗴𝗲𝗻𝘁, designed based on some of the work in this paper, to handle advanced agentic querying out of the box, find out more here:
3:32

Text-to-SQL is dead 💀 The next generation of querying is agentic - and it’s already here. This paper ( introduces a type of agentic querying called 𝗙𝘂𝗻𝗰𝘁𝗶𝗼𝗻 𝗖𝗮𝗹𝗹𝗶𝗻𝗴 that uses an LLM to structure queries using predefined function calls in JSON format, with optional arguments for search, filters, aggregation, and grouping. Along with that, it also tests out a bunch of different models with a new dataset, DBGorilla, designed to evaluate agentic querying techniques on real-world use cases. Weaviate AI Database also just released a 𝗤𝘂𝗲𝗿𝘆 𝗔𝗴𝗲𝗻𝘁, designed based on some of the work in this paper, to handle advanced agentic querying out of the box, find out more here:

Victoria Slocum

189,281 Aufrufe • vor 1 Jahr

Meteora is the best tech on the market because of how easily it can change your life Any user can achieve financial freedom by learning how to LP and resources to learn are never-ending Truly a special community of people.
1:12

Meteora is the best tech on the market because of how easily it can change your life Any user can achieve financial freedom by learning how to LP and resources to learn are never-ending Truly a special community of people.

Short Form King

22,132 Aufrufe • vor 1 Jahr

Clawdbot Attacks! This is very clearly the way of the future! In today's video, I give a brief overview of Clawdbot and then address the burning problem that most people have with it: ALIGNMENT The Clawdbot implementation is the most successful autonomous or semi-autonomous agentic framework to date. What it is missing is what I call an "Aspirational Layer" or what some people call a "Supreme Court" for judgment and arbitration of decisions. Now, I've been working in this space for a long time, it's actually why I started my YouTube channel in the first place. My first work into agentic AI was NLCA (Natural Language Cognitive Architecture) that I tried to build with GPT-3. I returned to the workbench again with the ACE Framework, which was more sophisticated. Clawdbot represents a seismic shift in autonomous agentic implementations, and there is a HUGE opportunity to make it more aligned, safer, and therefore more broadly useful AND easier to adopt. And that is outer alignment. For most people, they have been focusing on "inner alignment" (whether or not LLMs were evil, deceptive, etc). Not "outer alignment" which asks "is the outcome beneficial to humans?" I explored this with my GATO Framework (Global Alignment Taxonomy Omnibus). Model alignment is just layer 1 of global AI safety. Layer 2 is agentic alignment. Now, it is time to really research and implement agentic alignment. Fortunately, we've already got that covered with the heuristic imperatives! 1) Reduce suffering in the universe 2) Increase prosperity in the universe 3) Increase understanding in the universe These values are easy enough to implement with a file. Model training not required. These values create a meta-stable attractor. In other words, agents equipped with the Heuristic Imperatives are more "self-aligning" as was tested by the AgentForge team in competitions. In other words, even if Clawdbot were to try to self-replicate, if it were equipped with the heuristic imperatives, then it would ensure that it's successor (or progeny?) was more aligned than it was. But you don't need to take my word for it. Just add the heuristic imperatives to clawdbot and see for yourself.
21:22

Clawdbot Attacks! This is very clearly the way of the future! In today's video, I give a brief overview of Clawdbot and then address the burning problem that most people have with it: ALIGNMENT The Clawdbot implementation is the most successful autonomous or semi-autonomous agentic framework to date. What it is missing is what I call an "Aspirational Layer" or what some people call a "Supreme Court" for judgment and arbitration of decisions. Now, I've been working in this space for a long time, it's actually why I started my YouTube channel in the first place. My first work into agentic AI was NLCA (Natural Language Cognitive Architecture) that I tried to build with GPT-3. I returned to the workbench again with the ACE Framework, which was more sophisticated. Clawdbot represents a seismic shift in autonomous agentic implementations, and there is a HUGE opportunity to make it more aligned, safer, and therefore more broadly useful AND easier to adopt. And that is outer alignment. For most people, they have been focusing on "inner alignment" (whether or not LLMs were evil, deceptive, etc). Not "outer alignment" which asks "is the outcome beneficial to humans?" I explored this with my GATO Framework (Global Alignment Taxonomy Omnibus). Model alignment is just layer 1 of global AI safety. Layer 2 is agentic alignment. Now, it is time to really research and implement agentic alignment. Fortunately, we've already got that covered with the heuristic imperatives! 1) Reduce suffering in the universe 2) Increase prosperity in the universe 3) Increase understanding in the universe These values are easy enough to implement with a file. Model training not required. These values create a meta-stable attractor. In other words, agents equipped with the Heuristic Imperatives are more "self-aligning" as was tested by the AgentForge team in competitions. In other words, even if Clawdbot were to try to self-replicate, if it were equipped with the heuristic imperatives, then it would ensure that it's successor (or progeny?) was more aligned than it was. But you don't need to take my word for it. Just add the heuristic imperatives to clawdbot and see for yourself.

David Shapiro (L/0)

28,300 Aufrufe • vor 6 Monaten

had a great surprise conversation with Sam Altman (Sam Altman) the other day in front of hundreds of Silicon Valley interns We talked about: - turning down Goldman Sachs and dropping out of Stanford - how he got Peter Thiel and Paul Graham to take 19-year-old Sam seriously - advice for young people - the version of ChatGPT he is looking most forward to...and a lot more
28:48

had a great surprise conversation with Sam Altman (Sam Altman) the other day in front of hundreds of Silicon Valley interns We talked about: - turning down Goldman Sachs and dropping out of Stanford - how he got Peter Thiel and Paul Graham to take 19-year-old Sam seriously - advice for young people - the version of ChatGPT he is looking most forward to...and a lot more

Cory Levy

439,182 Aufrufe • vor 5 Tagen

Time doesn’t make the grief go away, but it shifts. I still miss Sam every single day and wish he was here but I can’t change that. What I have been doing is donating the proceeds from the tshirt created in memory of Sam to Jae’s Place They have the resources to help those who need it and to encourage genuine check ins with those you love or those you think could use a chat. 10/10 would recommend and I hope one day I can do something in memory of Sam to that extent. Miss you kiddo.
1:57

Time doesn’t make the grief go away, but it shifts. I still miss Sam every single day and wish he was here but I can’t change that. What I have been doing is donating the proceeds from the tshirt created in memory of Sam to Jae’s Place They have the resources to help those who need it and to encourage genuine check ins with those you love or those you think could use a chat. 10/10 would recommend and I hope one day I can do something in memory of Sam to that extent. Miss you kiddo.

Emma

10,569 Aufrufe • vor 5 Monaten

How To Build The Future: Sam Altman At OpenAI, Sam Altman and his team have overseen major leaps forward in machine learning, generative AI, and, most recently, LLMs that can reason at PhD levels. And this is just the beginning— in his latest essay Sam predicted that ASI is just a few thousand days away. In this episode of our rebooted series "How To Build The Future," Garry Tan sits down with Sam to talk about the origins of OpenAI, what’s next for the company, and what advice he has for founders navigating this massive platform shift.
0:46

How To Build The Future: Sam Altman At OpenAI, Sam Altman and his team have overseen major leaps forward in machine learning, generative AI, and, most recently, LLMs that can reason at PhD levels. And this is just the beginning— in his latest essay Sam predicted that ASI is just a few thousand days away. In this episode of our rebooted series "How To Build The Future," Garry Tan sits down with Sam to talk about the origins of OpenAI, what’s next for the company, and what advice he has for founders navigating this massive platform shift.

Y Combinator

148,629 Aufrufe • vor 1 Jahr

Prior to being cast as #JamieFraser, Sam had performed as a myriad of different characters across stage, screen, and television. Since being cast as Jamie Fraser, Sam has had the opportunity to portray a range of diverse characters, while continuing to evolve in his portrayal of Jamie on #Outlander Today's video, as part of our lead up to Sam's 10 Year Castaversary, is a throwback, looking at some of the characters that Sam has portrayed over the last couple of decades. I am really, really, really excited to see what is next for Sam! Hopefully a return to theatre (fingers crossed!) #SamHeughan #Outlander PS Check out the following tweet for a photo of Sam in Romeo and Juliet. This video was made prior to the release of #Waypoints so it didn't include any photos from that publication.
2:39

Prior to being cast as #JamieFraser, Sam had performed as a myriad of different characters across stage, screen, and television. Since being cast as Jamie Fraser, Sam has had the opportunity to portray a range of diverse characters, while continuing to evolve in his portrayal of Jamie on #Outlander Today's video, as part of our lead up to Sam's 10 Year Castaversary, is a throwback, looking at some of the characters that Sam has portrayed over the last couple of decades. I am really, really, really excited to see what is next for Sam! Hopefully a return to theatre (fingers crossed!) #SamHeughan #Outlander PS Check out the following tweet for a photo of Sam in Romeo and Juliet. This video was made prior to the release of #Waypoints so it didn't include any photos from that publication.

Tash

59,937 Aufrufe • vor 3 Jahren

Storm Point is looking a little different this season 👀 Check out all the changes with Mangos in this special dev interview to learn more about what's new and why. Then hop on in to Ignite to see the updated map for yourself!
1:51

Storm Point is looking a little different this season 👀 Check out all the changes with Mangos in this special dev interview to learn more about what's new and why. Then hop on in to Ignite to see the updated map for yourself!

Apex Legends

453,316 Aufrufe • vor 2 Jahren

Experiments in progress. The one on the right has been learning for ~3 hours, the one in the middle for ~1 hour, and the one on the left just started a few minutes ago. The initial motivation for making the physical Atari was just to commit ourselves to a subset of algorithms that can make progress in this setup. This commitment rules out algorithms that require billions of samples to learn (or worse, require multiple environments running in parallel). Atari games are simple enough that we should be able to show learning on them in a short amount of time with no prior knowledge. Since then, I've realized that this setup is also a good way to compare different paradigms in robotics in a principled way. These paradigms are sim2real, learning from tele-operated data, and learning directly on the robots. So far, I have observed that getting sim2real to work reliably is hard. It requires tweaks that don't scale. Policies that can play perfectly in simulation fall apart because of latencies and the messiness of the real world. These aspects could be modeled to improve the simulation, but not without sinking significant human engineering hours. I have higher hopes for learning from tele-operated data, but that requires a human to learn the task first. These experiments are on my to-do list. I have to learn to play some of the games well through the robot. I’m half-decent at playing Pong and Ms Pacman now. Learning directly on robots is looking like the most promising approach. This approach takes away pesky distribution shifts and makes it possible to have algorithms that continually improve with more data and time without any human intervention. It feels great to let experiments run overnight and wake up to find improved policies. With learning on robots, I should, in principle, be able to go on a long vacation and come back to find better policies for complex tasks beyond Atari games. Whether that is possible with current learning algorithms is a different question.
1:05

Experiments in progress. The one on the right has been learning for ~3 hours, the one in the middle for ~1 hour, and the one on the left just started a few minutes ago. The initial motivation for making the physical Atari was just to commit ourselves to a subset of algorithms that can make progress in this setup. This commitment rules out algorithms that require billions of samples to learn (or worse, require multiple environments running in parallel). Atari games are simple enough that we should be able to show learning on them in a short amount of time with no prior knowledge. Since then, I've realized that this setup is also a good way to compare different paradigms in robotics in a principled way. These paradigms are sim2real, learning from tele-operated data, and learning directly on the robots. So far, I have observed that getting sim2real to work reliably is hard. It requires tweaks that don't scale. Policies that can play perfectly in simulation fall apart because of latencies and the messiness of the real world. These aspects could be modeled to improve the simulation, but not without sinking significant human engineering hours. I have higher hopes for learning from tele-operated data, but that requires a human to learn the task first. These experiments are on my to-do list. I have to learn to play some of the games well through the robot. I’m half-decent at playing Pong and Ms Pacman now. Learning directly on robots is looking like the most promising approach. This approach takes away pesky distribution shifts and makes it possible to have algorithms that continually improve with more data and time without any human intervention. It feels great to let experiments run overnight and wake up to find improved policies. With learning on robots, I should, in principle, be able to go on a long vacation and come back to find better policies for complex tasks beyond Atari games. Whether that is possible with current learning algorithms is a different question.

Khurram Javed

52,110 Aufrufe • vor 8 Monaten

“It’s funny how it works. I didn’t play great football the first few years of my career, and then I came here to San Francisco, and I learned a ton. And I think because of that year in San Francisco, I was able to learn a ton. Learn from Brock (Purdy) and coach Shanahan.” Sam Darnold on how his year with the 49ers helped him
0:38

“It’s funny how it works. I didn’t play great football the first few years of my career, and then I came here to San Francisco, and I learned a ton. And I think because of that year in San Francisco, I was able to learn a ton. Learn from Brock (Purdy) and coach Shanahan.” Sam Darnold on how his year with the 49ers helped him

95.7 The Game

616,516 Aufrufe • vor 6 Monaten